ML Engineer

About Nebius:

Nebius is leading a new era in cloud infrastructure for the global AI economy. We are building a full-stack AI cloud platform that supports developers and enterprises from data and model training through to production deployment, without the cost and complexity of building large in-house AI/ML infrastructure.

Built by engineers, for engineers. From large-scale GPU orchestration to inference optimization, we own the hard problems across compute, storage, networking and applied AI.

Listed on Nasdaq (NBIS) and headquartered in Amsterdam, we have a global footprint with R&D hubs across Europe, the UK, North America and Israel. Our team of 1,500+ includes hundreds of engineers with deep expertise across hardware, software and AI R&D.

The role

This role is for Nebius AI R&D, a team focused on applied research in AI. Examples of applied research that we have recently published include:

• applying reinforcement learning for agent training in long-context multi-turn scenarios
• dramatically scaling task data collection to power reinforcement learning for SWE agents
• building a decontaminated evaluation for SWE agents that is regularly updated
• investigating how test-time guided search can be used to build more powerful agents

The results often lead to collaboration with adjacent teams where our research findings are applied in practice.

We are currently looking for senior- and staff-level ML engineers to work on research in areas such as:

• Guided search and reinforcement learning for agentic systems
• Reinforcement learning for reasoning models
• Web-scale problem collection for training agents
• Efficient model distillation

Some examples of what your responsibilities might include are:

• Conducting experiments to figure out efficient ways to train a large language model on traces of interactions with various environments
• Exploring methods of guided generation and search in the trajectory space
• Coming up with ways to mine relevant data at web scale and figuring out efficient ways to use this data in model post-training
• Conducting experiments with different reinforcement learning configurations in verifiable domains